Comparison of the Charlson Comorbidity Index derived from self-report and medical record review in Asian patients with rheumatic diseases

The aim of the study was to compare the agreement between self-report Charlson Comorbidity Index (SR-CCI) and the medical record-based CCI (MR-CCI) and to examine the impact of both instruments on health-related quality of life (HRQoL) amongst Asian patients with rheumatic diseases. This cross-sectional study surveyed a convenience sample of patients seen at rheumatology specialty outpatient clinics. Patients completed the SR-CCI and Short Form 36, while two research assistants completed the MR-CCI. Item-level agreement between the SR-CCI and MR-CCI was evaluated using kappa coefficients. Adjusted linear regression models evaluated the independent effect of the SR-CCI/MR-CCI on HRQoL. The study included 301 patients (median age 51, range 21-79, 61.5 % female, 68.8 % Chinese, 17.6 % Indian, 6.0 % Malay). Kappa statistics for cerebrovascular disease (0.433), chronic pulmonary disease (0.509), connective tissue disease/rheumatoid arthritis (0.506), ulcer disease (0.461), and tumour (0.541) reflected moderate agreement between the SR-CCI and MR-CCI (all p < 0.0001). There was substantial agreement in the reporting of diabetes (0.764, p < 0.0001) but poor/fair agreement for that of myocardial infarction (0.359, p < 0.0001) and diabetes with end-organ damage (0.189, p = 0.0002). Increases in SR-CCI were associated with significant reductions in both physical (beta coefficient -2.56, p < 0.0001) and mental HRQoL (beta coefficient -1.24, p = 0.044). However, such associations were not observed with the MR-CCI. The SR-CCI demonstrated moderate concordance with the MR-CCI, and the SR-CCI but not MR-CCI scores were associated with lower HRQoL. Assessment of comorbidities amongst rheumatology patients remains complex, and more efficient methods of quantifying these conditions are needed for clinical and research purposes.